结合LangChain实现网页数据爬取

# 定义大模型from langchain_openai import ChatOpenAIllm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo-0613")
# 定义提取方法def extract(content: str, schema: dict):from langchain.chains import create_extraction_chainreturn create_extraction_chain(schema=schema, llm=llm).invoke(content)
import pprintfrom langchain_text_splitters import RecursiveCharacterTextSplitterdef scrape_with_playwright(urls, schema):# 加载数据loader = AsyncChromiumLoader(urls)docs = loader.load()# 数据转换bs_transformer = BeautifulSoupTransformer()# 提取其中的span标签docs_transformed = bs_transformer.transform_documents(docs, tags_to_extract=["span"])# 数据切分splitter = RecursiveCharacterTextSplitter.from_tiktoken_encoder(chunk_size=1000, chunk_overlap=0)splits = splitter.split_documents(docs_transformed)# 因为数据量太大，输入第一片数据使用，传入使用的架构extracted_content = extract(schema=schema, content=splits[0].page_content)pprint.pprint(extracted_content)return extracted_content
urls = ["https://ceshiren.com/"]schema = {"properties": {"title": {"type": "string"},"url": {"type": "string"},},"required": ["title", "url"],}extracted_content = scrape_with_playwright(urls, schema=schema)